Identify highest probability in regression analysis?

I was studying the matlab example for logistic regression linked below and under the heading "Evaluating the Model Predictions" I m trying to identify what the value .95 is for the line of code

 [failedPred,dlo,dhi] = glmval(logitCoef,weightPred,'logit',stats,.95,100);

It appears as though its the highest probability from the normplot but I don't know how this was obtained. Looking at the help for glmval I m still not sure what it means. If anybody at matlab knows, I d like appreciate an answer.

http://www.mathworks.com/examples/statistics/2142-fitting-data-with-generalized-linear-models